[Gtkradiant] mo' bullshit
Fri, 29 Mar 2002 17:41:15 +1100
Maybe we should take advantage of the exposure this rant/diary/whatever is
getting (linked from a couple of news sites), to up the level of feedback we're
getting. Perhaps declare a "complain about radiant day" or "bug week" (with a
cooler name obviously ;) or something to encourage people that are noticing
bugs to report them when they wouldn't usually. This could also be tied into the
releases that we do (i seem to remember doing release candidates more often
than we do now).
Using the next stable release (1.2.7) as an example;
-> Release 1.2.7 as rc/update/nightly/testing/whatever
Encourage everyone to use/test it and report bugs stating explicitly that it is
not a stable release, however stable it is, maybe simplifying the bug reporting
process could help here also. Once the 1.2.7 version is released, no features
are added, only bug fixes.
-> Release 1.2.8 as stable
Begin adding features, and fixing reported bugs for the 1.2.9 release.
-> rinse & repeat, version++;
This method is similar to linux kernel releases, in some respects, where odd
numbered releases are considered test releases, and even numbers, stable.
I think it's also important not to jump on people who are reporting duplicate
bugs, which would discourage them imho... Atleast they're making the effort
to report them to begin with, and that needs to be encouraged. Maybe adding
who reported the bug to the changelog, or the new website would encourage
them in some way. A bug reporting plugin could be something to look at.
Also I agree with ydnar about the package size, 20mb+ (rad+q3a+wolf) is a
touch large *g*. And there haven't been any major changes to the q3a media
for a long time now, so redistributing that with each major release shouldn't
really be necessary. The down side to breaking it up into smaller modules is
that it may be confusing for new mappers.
Anyway, just some thoughts.
On Thu, 28 Mar 2002 10:16:13 +0100
Timothee Besset <firstname.lastname@example.org> wrote:
> I think a bunch of people here have already read through this:
> We are pretty used to that kind of rant, it's not really a problem
> anymore. However it seems it's that time of the year were we are going to
> need to explain the usual things about Radiant again, how it evolves,
> where does it come from, how QA happens in an open source project etc.
> With the advent of HL/CS support and more multiple games support (who said
> JKII?), QA is going to become even more tricky, and we're gonna get even
> more public exposure (specially to newbies), so I say we should expect to
> see more and more of that.
> I feel that there is one true problem though in the way the releases have
> been evolving. Stable release / nightly updates and stuff .. isn't that
> clear anymore .. even for me. We came from a situation were we had 1.1 or
> 1.1-TA as the stable release, and 1.2 were only nightly releases. At that
> time, nightlies were truly beta quality, and we recommended 1.1* as the
> production editor.
> Now we have completely shifted focus. There is no build of 1.3 available
> at this time, and the 1.2 nightlies are actually stable updates over the
> last 1.2.3 stable release. I still want to call them nightlies because
> they are meant as 'release often' and 'small updates over existing
> install'. And I won't call them beta either, because they truly are stable
> updates over something we already tagged as 'stable'.
> We did such a good job telling people that nightlies were experimental
> when we first introduced them, that they haven't recovered yet and are
> reluctant to use 1.2.6
> I'm not sure what kind of action we should take. Could just post on
> qeradiant.com with some more explanations. But I think a more definitive
> page on the website may be good. Some kind of manifest, explaining how the
> QA is done, how the releases are handled and stuff.
> feedback welcome obviously :)
> Gtkradiant mailing list